Notice: On April 23, 2014, Statalist moved from an email list to a forum, based at statalist.org.
| From | Tanuku AP <tanuku.ap@hotmail.com> |
| To | <statalist@hsphsun2.harvard.edu> |
| Subject | st: probit - interaction mfx when dummy has >2 categories |
| Date | Sat, 9 Jul 2011 19:22:02 -0400 |
Hello, Is there a ready way to deal with interactions (in probit models) when dummy variable has more than 2 categories? a) find out the correct marginal effects b) test if interactions are jointly significant I have quickly read through --inteff-- [Ai and Norton] but it seems that --inteff-- deals with binary qualitative variables only. -------------------------------------------------------------- For my problem, we have binary variable 'disabled' and five income brackets: income1 - income5. We want to test for joint significance and find out the interaction marginal effects (if any). We cannot collapse the income brackets into 2 categories. > gen income1d = income1 * disabled //interaction term .... > logit choice disabled income1 income2 income3 income4 income1d income2d income3d income4d ... test income1d income2d income3d income4d --Tanuku * * For searches and help try: * http://www.stata.com/help.cgi?search * http://www.stata.com/support/statalist/faq * http://www.ats.ucla.edu/stat/stata/